© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129108
    Prev
    Next
    INDEX
    Explanations

    This neuron fires on the first substantive content words at the start of a new section or paragraph—that is, the initial contentful tokens that introduce each segment’s topic.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    items for sale and places

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     now
    -1.08
    rangle
    -0.99
    y
    -0.96
    하게
    -0.95
     metus
    -0.93
     sinto
    -0.93
    )
    -0.91
     asfalto
    -0.90
    ших
    -0.90
     as
    -0.90
    POSITIVE LOGITS
     available
    1.53
     offered
    1.32
     proposé
    1.30
     options
    1.25
     selection
    1.23
     представлены
    1.21
     browsing
    1.21
    offered
    1.20
     καλύτε
    1.18
     agricultores
    1.18
    Activations Density 0.042%

    No Known Activations